Film

Evaluation of compensated heat-pulse velocity method to determine vine transpiration using combined measurements of eddy covariance system and microlysimeters

Description
Evaluation of compensated heat-pulse velocity method to determine vine transpiration using combined measurements of eddy covariance system and microlysimeters
Categories
Published
of 9
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
  AgriculturalWaterManagement 109 (2012) 11–19 ContentslistsavailableatSciVerseScienceDirect Agricultural   Water   Management  j   ournal   homepage:www.elsevier.com/locate/agwat Evaluation   of    compensated   heat-pulse   velocity   method   to   determine   vinetranspiration   using   combined   measurements   of    eddy   covariance   system   andmicrolysimeters C.Poblete-Echeverría a , ∗ ,   S.   Ortega-Farias a ,M.   Zu˜niga a ,S.Fuentes b a ResearchandExtensionCenterfor    IrrigationandAgroclimatology(CITRA),UniversidaddeTalca,Talca,Chile b The   Universityof     Adelaide,SchoolofAgriculture,FoodandWine,PlantResearchCenter,WaiteCampus,PMB1,GlenOsmond,S.A.5064,Australia a   r   t   i   c   le   i   nf   o  Articlehistory: Received24August2011Accepted31January2012 Available online 3 March 2012 Keywords: SapflowCompensatedheat-pulsevelocityEddycovariancesystemMicrolysimetersMerlotvineyard a   b   s   t   ra   ct A   fieldexperiment   wascarried   out   with   the   objectivetoevaluate   the   compensated   heat-pulse   velocity(CHPV)   methodused   to   determinevinetranspiration(Tsap).   Theperformance   ofthe   CHPV   methodwasevaluated   usingdailyvalues   of    residual   transpiration   (Tr),obtained   as   the   difference   between   actualevapotranspiration   (ETa)   and   soilevaporation   (Es)(Tr=   ETa   −   Es)measured   from   an   eddy   covariance   (EC)systemandmicrolysimeters,   respectively.   Data   usedinthis   study   were   collected   overa   drip-irrigatedMerlot   vineyard   trainedonavertical   shoot   positioned(VSP)systemduringthree   consecutive   growingseasons(2006/2007,2007/2008and2008/2009).Results   showed   that   the   bestagreement   betweenTsapand   Trwasobtained   usingcorrectioncoefficients   for   awoundsize   of    2.4mm.   Thecomparison   betweenTsap   andTrindicated   that   the   index   of    agreement( d )was   0.97,androotmean   square   error   (RMSE),   meanabsolute   error   (MAE)   and   mean   bias   error   (MBE)   were   0.22,   0.18and   − 0.04mmday − 1 ,respectively.   Also,the   sensitivityanalysis   of    fraction   of    wood(FM),   fractionof    water   (FL)andfactorof    thermal   propertiesof    the   woody   matrix   ( k )suggested   that   the   changesof    ± 30%have   alittle   effectinthe   finalestimation   of dailyvinetranspiration   with   variations   lessthan   12%.   Finally,   major   disagreements   betweenTrand   Tsapwere   observed   onpartially   cloudydays   where   rapid   changes   (on   30mintimeintervals)   of    solarradiationproduced   extreme   values   of    volumetricsap   fluxdensity. © 2012 Elsevier B.V. All rights reserved. 1.Introduction Toachieveanoptimalgrapevine( Vitisvinifera L.)productionindrip-irrigatedvineyards,itisnecessaryanaccurateestimationof theactualevapotranspirationcomponents(ETa)[i.e.vinetranspi-rationandsoilevaporation]to   assessvineyardwaterrequirements(Trambouzeetal.,   1998;Yunusaetal.,2004).However,theesti-mationofETacomponentsinvineyardsis   a   complextaskduetodiscontinuouscanopiesandarchitectureimposedbytrellissys-tems(Heilmanetal.,1994;TrambouzeandVoltz,2001;Yunusaetal.,2004).Vineyardstrainedonverticalshootpositionedsys-tem(VSP)presentalowfractionalcover(  f  c  valuesabout20–40%),whichproduceconsiderablevariabilityinsolarradiationexposureofcanopiesandsoilbetweenrows(Heilmanetal.,   1996;Ortega-Fariasetal.,2007).Furthermore,indrip-irrigatedvineyards,thewettedfractionofsoilsurfacebydrippersis   low,reducingconsid-erablysoilevaporation(Lascanoetal.,   1992;Ortega-Fariasetal., ∗ Correspondingauthor.Tel.:+5671   200426;fax:+5671   200212. E-mail   addresses: cpoblete@utalca.cl,poblete.cl@gmail.com(C.Poblete-Echeverría). 2010).Therefore,thequantificationofvinetranspirationin   drip-irrigatedvineyardsunderdry   atmosphericconditionsishighlyimportant,sincerepresentsbetween70%and80%of    ETa(Ortega-Fariasetal.,2008).Currently,severaltechniquesareavailableto   measuretran-spirationin   plants(e.g.lysimeters,volumetricmethodsandgascanopyanalyzers).However,thepracticalapplicationof    themajor-ityofthesetechniquesislimitedbyitscost,andrepresentativeness(Dugasetal.,1993;SmithandAllen,1996).Thus,in   thelast15yearssapflowsensorshavebecomethemostusedmethodto   esti-matewhole-planttranspirationunderfieldconditionsforresearchpurposes.Thevinetranspirationcanbedeterminedusingheatsystems,whichuseheatasa   tracerrelatedtowaterflowthroughthesap.Thesesystemscanbe   groupedmainlyinthreetypes(González-Altozanoetal.,   2008):(i)heat-pulse(e.g.EasthamandGray,1998;Yunusaetal.,   2004;Zhangetal.,2009),(ii)stemheatbalance(e.g.BraunandSchmid,1999a;Trambouzeetal.,1998;TrambouzeandVoltz,2001)   and(iii)thermalheatdissipation(e.g.BraunandSchmid,1999b;Luetal.,   2003;Patakasetal.,2005).Theuseof    sapflowsensorsinvineyardshasbeenappliedmainlyto:(i)irrigationmanagementand(ii)calibrationandvalidation 0378-3774/$–seefrontmatter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.agwat.2012.01.019  12  C.Poblete-Echeverríaetal./AgriculturalWaterManagement  109 (2012) 11–19 oftranspirationmodels.Forirrigationmanagement,someauthorshaveevaluatedtheuseof    sapflowsensorstoestimatevinetranspi-rationondifferentgrapevinecultivarsandatmosphericconditions.Ginestaretal.(1998)andEasthamandGray(1998)indicatedthat sap-flowsensorscanprovideusefulinformationforirrigationman-agementofvineyards.Thus,differentirrigationstrategiessuchasregulateddeficitirrigation(RDI)usedto   controlvigorandoptimizegrapequalitycanbeappliedbasedonvinetranspirationmeasuredbysapflowsensors,especiallyfordrip-irrigatedvineyards.Othertypesofstudieshavebeencarriedoutwiththeobjectivetocalibrateandvalidatetranspirationmodelsovervineyardssuchas:(i)theuseofGraniersapflowsensorsto   evaluatea   transpirationmodelinnon-waterstressedvinescv.Sultana(Luetal.,2003),(ii) theuseofthecompensatedheat-pulsevelocity(CHPV)methodinadrip-irrigatedMerlotvineyardtodevelopa   simplemodeltoesti-matevinetranspirationbasedonthePenman–Monteithapproach(Pereiraetal.,2006)and(iii)theuseofheat-pulsesapflowsensors tovalidateamodelbasedon   thetwolayermodelofShuttleworthandWallacetoestimatevinetranspirationin   a   vineyard(cv.Mer-lot)underpartialroot-zonedrying(Zhangetal.,2009).Allofsapflowtechniques,mentionedpreviouslyhavethepotentialtoesti-matetranspirationin   vineyards.However,heat-pulsetechniquesareoftenpreferredbecauseoftheirsimpleinstrumentation,andlowpowerrequirements(Swanson,1994;SmithandAllen,1996). Also,heat-pulsetechniquesallowstomeasuresapflowtakingintoaccountthevariabilityofthexyleminthecrosssectionthroughtheuseofthermocouplesatdifferentdepth(Greenetal.,2003;González-Altozanoetal.,2008).Inthisregard,CHPVmethodhasbeenwidelyusedinfruittreessuchas:(i)olivetrees( Oleaeuropaea L.)   (e.g.Morenoetal.,1996;Fernándezetal.,   2001,2003,2006;GiorioandGiorio,2003;Williamsetal.,2004;Tognettietal.,2004;Pereiraetal.,   2006,2007),(ii)peachtrees( Prunuspersica L.)   (e.g.Conejeroetal.,2007;González-Altozanoetal.,2008),(iii)appletrees( Malusdomestica L.)(e.g.Greenetal.,   1989,2003;Gongetal.,2007;Pereiraetal.,2006,2007),andgrapevines( V.   vinifera )   (e.g.EasthamandGray1998;Ginestaretal.,1998;Yunusaetal.,2004;Pereiraetal.,2006;Zhangetal.,2009).Also,González-Altozanoetal.(2008)in   peachtreesindicatedthatCHPVmethodwasthemostsensitivesystemindetectingwaterstressunderexperimentalconditions.Thischar-acteristicisveryimportantincommercialvineyards,whicharemanagedunderRDIforoptimizingthewateruseandgrapequality.However,severalreportspointedoutuncertaintiesin   heat-pulsetechniques,suchastheinvasivenatureof    sapflowsensors,whichrequiretheuseofsemi-empiricalcoefficientstocorrectheat-pulsevelocity( V  )(Greenetal.,   2003).Also,in   theformulationof sapflowcriticalassumptionsaremadeintheestimationof    thefol-lowingparameters:(i)sapfluxdensity(  J  )(determinationofwood(FM)andwater(FL)volumefractionsandfactorof    thermalproper-tiesofthewoodymatrix( k )factor),(ii)volumetricsapflow( Q  )(determinationof    sapwoodconductingarea)and(iii)transpira-tion(scalingupprocedure)(ˇ Cermáketal.,2004;Davidetal.,2004;Dragonietal.,2005;Crosbieetal.,2007).Therefore,itis   question-abletoassumethatsapflowsensorsareaccurateunlesstheyaretestedwithanalternativetechnique.Inthisregard,Dragonietal.(2005)evaluatedtheCHPVmethodusingasa   referencetheappletranspirationmeasuredfromcanopygasexchangechambers.Thisstudyindicatedthatitisnecessarytocalibratesapflowmeasure-mentstoobtainreliableestimatesof    transpirationrates.Inthecaseofgrapevines,Intriglioloetal.(2009)usingheat-pulsesapflowsensorswiththeTmaxmethodindicatedthatthedirectsapflowtranspirationwasnot   reliablebecausethedifferencesbetweentranspirationobtainedbysapflowsensorsandthatmeasuredbythecanopygasexchangechamberswereconsiderableandvariedfromvinetovine.Thisstudysuggestedthattherewas   nota   sys-tematicerrorinsapflowreadingsbutratheraseeminglyrandomdeviationfromthereferencevalues.Therefore,themainobjectiveofthisresearchwas   toevaluatetheuseofCHPVmethodtoobtaindailyvinetranspirationfora   drip-irrigatedMerlotvineyardtrainedonVSPundersemi-aridconditions.ThestandardvalueusedforevaluatingCHPVmethodwastheresidualtranspirationobtainedasthedifferencebetweenETaandEswhichweremeasuredusingtheeddycovariancesystemandmicrolysimeters,respectively.Addi-tionally,a   sensitivityanalysiswasperformedtoevaluatetheeffectofwoundsizes,FM,   FL    and k intheestimationof    TsapusingtheCHPVmethod. 2.Materialsandmethods  2.1.Studysite Thefieldexperimentalplotwas   locatedin   theTalcaValley,MauleRegion,Chile(35 ◦ 25 ′ LS;71 ◦ 32 ′ LW;   125m.a.s.l.)duringthreeconsecutivegrowingseasons(2006/2007,2007/2008and2008/2009).TheclimateisMediterraneansemi-aridwithanaver-agedailytemperatureof17.1 ◦ C   anda   meanannualrainfallof 679mm.   Thesummerperiodisusuallydryandhot(2.2%of    annualrainfall)whilethespringiswet   (16%of    annualrainfall).ThesoilatthevineyardisclassifiedasTalcaseries(Fine,mixed,thermicUlticHaploxeralfs)withaclayloamtextureandanaveragebulkdensityof    1.5gcm − 3 .At   theeffectiverootingdepth(0–60cm),thevolumetricsoilwatercontentatfieldcapacityandwiltingpointwere0.36m 3 m − 3 and0.22m 3 m − 3 ,   respectively.Thevineswereirrigateddailyusing4L    h − 1 drippersspacedatintervalsof1.5m.Merlotvinesgraftedon101–14Mgt.(rootstockusedtocontrolvigor)wereplantedin   1999innorth–southrowswitha   distancebetweenrowsequalto   2.5m,   adistancewithinrowsof    1.5mandweretrainedonVSPwiththemainwire1mabovethesoilsur-face.Shootsweremaintainedinaverticalplanebythreewires,thehighestonewas   located2mabovethesoilsurface.  2.2.Complementaryfieldmeasurements Irrigationswerescheduledconsideringa   maximumalloweddepletionof    waterfromsoilprofileof    29%ateffectiveroot-zonedepth(0.6m).   Thevolumetricsoil   watercontentattheroot-zonedepthwas   monitoredweeklyin12samplingpointsdistributedinsidethevineyardusinga   portableTimeDomainReflectometryunit(TRASE,SoilMoistureCorp.,SantaBarbara,California,USA).Vine   waterstatuswas   evaluatedweeklyusingthemiddaystemwaterpotential(   x )measuredbya   pressurechamber(PMS600,PMS   InstrumentsCompany,Corvallis,Oregon,USA).   x  wasmea-suredon12youngfullyexpandedleaves(2leavespervine).Theleaveswereselectedfromthemiddlezoneofthecanopyandwerebaggedontheshootsin   plasticbagscoatedwithaluminumfoilforatleast2hbeforemeasurements(Chonéetal.,   2001).Themea-surementswithpressurechamberweremadeimmediatelyaftercuttingleavesfromtheshoots.Thisparameterwasusedtoevalu-atethelevelof    waterstressofthevineyard(WilliamsandTrout,2005;Sibilleetal.,   2007).Theaverage  f  c  ofthevineyardwas   estimated10timesduringeachseasonbymeasuringtheprojectedareaoccupiedbythevine(fractionof    groundshadedbytheverticalprojectionof    thecanopyatmidday),usingdigitalimages.Theanalysisofthedigitalimages(whiteandblackpixelcount)wasmadeusingascriptwritteninMATLAB ® 2009a(TheMathworksInc.,Natick,MA,   USA).  2.3.Meteorologicalandmicrometeorologicalmeasurements Referenceevapotranspiration(ETo)wascalculatedbyPenman–Monteithequation(FAO-56Method)(Allenetal.,1998)usingdailymeteorologicalvariablesobtainedbya   nearby  C.Poblete-Echeverríaetal./AgriculturalWaterManagement  109 (2012) 11–19 13 weatherstation“Panguilemo”(35 ◦ 37 ′ LS;71 ◦ 59 ′ LW;   118ma.s.l.).Also,weathervariablessuchasair   temperature(Ta)andrelativehumidity(RH)weremeasuredwithintheexperimentalvineyardplotusinga   Vaisalaprobe(HMP45C,CampbellScientificInc.,Logan,Utah,USA).Windspeed(Ws)andwinddirectionweremeasuredusingacupanemometerandawindvane(YOUNG,03101-5,Michigan,USA),respectively.Solarradiation(SR)wasmeasuredwithaSiliconPyranometer(LI200X,CampbellScientificInc.,Logan,Utah,USA)andprecipitationwasmeasuredwithapluviometer(RGB1,CampbellScientificInc.,Logan,Utah,USA).Allvariablesweremonitoredat4.7mfromthesoilsurface.Vaporpressuredeficit(VPD)wascalculatedfromhourlyaveragesof    TaandHRaccordingto   Allenetal.(1998).Soilheatflux( G )wasmeasuredbyheatflux   platesofconstantthermalconductivity(HFT3,CampbellScientificInc.,Logan,UT,USA).Netradiationoverthevineyard(Rn)   wasmeasuredat4.7   mbyafour-waynetradiometer(CNR1,Kipp&ZonenInc.,Delftechpark,TheNetherlands).Fluxesof    latent(LE)andsensible( H  )heatweremeasuredusinganeddycovariance(EC)system,whichconsistedof    afastresponseopen-pathinfraredgasanalyzerIRGA(LI-7500,LI-CORInc.,Lincoln,Nebraska,USA)anda   three-dimensionsonicanemometer(CSAT3,CampbellScientificInc.,Logan,UT,USA).Moredetailsaboutthedistributionandconfigurationofthesen-sorsusedtomeasurethevineyardenergybalancecomponents( G ,Rn, H  andLE)arepresentedinPoblete-EcheverriaandOrtega-Farias(2009).ECsystemconsistsbasicallyina   high-frequencymeasure-mentofwindandscalaratmosphericdataseries(inthisstudyheatandwaterfluxes).Theaccuracyof    theECsystemmeasurementswasevaluatedusingtheenergybalanceratio(EBR)calculatedonadailybasisasfollows:EBR  = H  + LERn −   G  (1)where H  ,LE,Rn   and G areexpressedin   MJm 2 day − 1 andEBRistheenergybalanceratio(dimensionless).Dayspresentingerrorslessthanof10%(i.e.EBR>0.9orEBR<1.1)in   theclosurewereusedinthisstudy.  2.4.Soilevaporation Soilevaporationwasmeasuredweeklyusingcylindricalmicrolysimetersduringrainlessperiods:(a)11daysforfirstseason(2006–2007);(b)14daysforthesecondseason(2007–2008);and(c)   8daysforthethirdseason(2008–2009)(BoastandRobertson,1982;Trambouzeetal.,   1998;Yunusaetal.,2004).Asetofsixmicrolysimeterswereinstalledtakingintoaccountthedifferentconditionspresentedwithinthevineyard,suchas:shaded,sun-lit,wettedanddryzones.Twomicrolysimeterswereinstalledinthewettedzonenextto   thedrippersunderthecanopy,andfourmicrolysimeterswereinstalledin   thedryzonebetweenrows,ofwhichtwowereinstalledinwestsideandtwoin   theeastside,toconsiderthemovementoftheshadethroughtheday-time.MicrolysimetersweremadefromPVCtubeswithaninternaldiameterof75mm   anda   depthof150mm(Yunusaetal.,   2004).Soilevaporationateachmicrolysimeterswasobtainedasthedif-ferencebetweentheweightsmeasuredat8:00and20:00h(localtime).Thiswasdoneusinganelectronicscale(BB2440Deltarange,Mettler-Toledo,Greifensee,Switzerland)witha   resolutionof    0.01ginaflatsurfaceinsideoftheirrigation-house.Finally,dailysoilevaporationofthevineyard(Es)(mm   day − 1 )wascalculatedasfol-lows:Es = (1 −  f  c )Ebr + (  f  c )Ec(2)whereEbristhesoilevaporationbetweenrow(mmday − 1 )andEcisthesoilevaporationbelowthecanopynext   to   thedrippers(mmday − 1 ).  2.5.Sapflowmeasurements Thevinetranspirationwas   determinedcontinuallyusingsapflowsensors(700SF-100,Tranzflo,NewZealand)insertedinthetrunkof    eightvines(onesapflowsensorspervine)(Fig.1).Homo-geneousvinesnearto   theECsystemwereselectedwithsimilarstemdiameters,rangingfrom113to   124mmin   thefirstseason110to126mm   in   thesecondseasonand112to   122mminthethirdseason.InthisstudytheCHPVmethodwasusedto   mea-surevinetranspiration.Thismethodusestwotemperaturesensorprobesinstalledasymmetricallyaboveandbelowa   lineheaterprobe(Greenetal.,2003).Theupstreamsensorprobeislocated5mm   belowtheheaterprobeandthedownstreamsensoris   located10mm   abovetheheaterprobe.Oneachsensorthereare   threether-mistors.Thethreethermistorsforbothupstreamanddownstreamsensorsare   positionedat5mm,   10mm   and15mm.   Thethermistorsarepairedona   verticalplanetofacilitatethemeasurementofsapflowvelocity.A   metallicplatewithpre-drilledholeswas   usedasaguidetodrillthethreeholesandtoassurea   paralleldisplacementof theprobeneedles;drillbitsof1.8mmin   diameterwereusedto   drilltheholes(Fig.1a).Afterinsertion,theprobeswerecoveredwithaluminumfoiltoreducetheeffectsof    ambienttemperatureandsolarradiation(Fig.1b).Adatalogger(CR23X,CampbellScientificInc.,Logan,UT,USA)wasusedtotriggerheat-pulsesandrecordthetimeintervalbetweeninitiationof    a   heat-pulseandequilibrationof    thetemperaturedifferencebetweenupstreamanddownstreamtemperatureprobes( t  z ).Heat-pulsesof0.75swereautomaticallytriggeredeveryhourthroughouttheday.Thesystemwaspoweredbyacidbatteriesconstantlyrechargedbysolarpanels.  2.6.FormulationofCHPVmethod Heat-pulsesystemsrelyontheinsertionof    probescontainingsensorsin   thexylemtomeasuretemperaturealterationsproducedbya   heat-pulse(González-Altozanoetal.,2008).Inthisstudyrawheat-pulsevelocity( V  )(cmh − 1 )wascalculatedasfollows(Greenetal.,2003): V  =   x d +  x u 2 t  z  3600(3)where t  z is   thetimetothermalequilibrationof    thedownstreamandupstreamprobesafterreleaseof    theheat-pulse(s),   3600is   atimeconversionfactor,and  x d  and  x u  denotedistances(cm)betweentheheaterandthedownstreamandupstreamtemperatureprobes,respectively.Anegativevalueis   assignedto  x u  becauseitislocatedontheoppositeside   of    theheater.Theinstallationofprobesinthestemcausesa   disturbanceof    thenaturalxylemconditionsandconsequently, V  valuesmustbecorrectedfortheprobeinducedeffectsof    woundingusinganequationasfollows:Vc = a 0 + a 1 V    + a 2 V  2 (4)whereVcis   thecorrectedheat-pulsevelocity(cmh − 1 ). a 0 ,   a 1 ,   and a 2  arecorrectioncoefficientsforwoundsizes.Inthisstudy,thewoundsizeusedwas   2.4mm,   so   a 0 , a 1 ,   and a 2  coefficientswere0.394,1.12,and0.0878,respectively(Greenetal.,2003).Thenextstepis   tocalculatesapfluxdensity,(  J  )(cmh − 1 ),asfollows:  J  = ( k FM   + FL )Vc(5)where  J    isthesapfluxdensityFMandFLarethevolumefractionsof woodandwater,respectively.Thefactor k   isrelatedto   thethermalpropertiesofthewoodymatrix,anditis   assumedtobeconstant( k =0.441)withinandbetweenspecies(BeckerandEdwards,1999).ThevolumefractionsFMandFLweredeterminedusingvaluesof wooddensityandmoisturecontentdeterminedin   27woodcoresamplestakenfromoneplantof    thestudysite.TheaveragevaluesofFMandFLobtainedwere0.46and0.52,respectively.  14  C.Poblete-Echeverríaetal./AgriculturalWaterManagement  109 (2012) 11–19 Fig.1. Photographyofthesapflowsensorinstallation,showingthesensorposition(a)   andthealuminumfoilcover(b). Avolumetricdeterminationof    totalsapflowinthevinewasobtainedusingthesapfluxdensitiesatthethreedifferentpointsoverthetotalsapwoodconductingareaasfollows(Hattonetal.,1990): Q  = n  i = 1  A n  J  n 1000 (6)where Q  isthevolumetricsapflow(L    h − 1 ),1000isaconversionfactorforcm 3 toL,  J  n  arethesapflux   densitiesatthreesapwoodconductingareas, n isthenumberof    thermistorsforeachsensorprobe( n =3;at5,10and15mmdepth)and  A n  arethedifferentsapwoodconductingarea(cm 2 ).Finally,dailyvinetranspirationestimatedbysapflowsensors(Tsap)(mm   day − 1 ),was   calculatedby   thefollowingexpression:Tsap =  j  i = 1 Q  3 . 75 (7)where3.75isthedesignatedplantingareapervine(m 2 )and  j   isthenumberofhourlymeasurementsinaday(  j =   24).  2.7.EvaluationofCHPVmethod ToevaluatedthevinetranspirationobtainedbyCHPVmethod,theresidualtranspirationcalculatedasthedifferencebetweenETaandEs(Tr=ETa − Es)was   theuseasstandardvalue.Theevalu-ationincludedalinearregressionanalysisbetweenTr   andTsapandthe t  -testtoevaluatethenull   hypothesis:interceptequaltozeroandslopeequalto   unityatthe95.0%confidencelevel(i.e. ˛ =0.05).Also,therootmeansquareerror(RMSE),meanabsoluteerror(MAE),meanbiaserror(MBE),indexof    agreement( d )andmodelefficiency(EF)werecalculated(Willmott,1982;MayerandButler,1993;LegatesandMcCabe,1999)(Table1).Additionally,a sensitivityanalysiswasperformedto   evaluatetheeffectof    woundsizes,FM,   FLand k   intheestimationof    TsapusingtheCHPVmethod.Inthiscase,valuesof    FM,   FLand k werevariedby ± 30%.Fordeter-miningsensitivityofFM,   FL    and k   wecalculateda   relativechangeexpressedasapercentage.Also,itincludedthesensitiveindex(SI),whichcorrespondstotheoutputpercentagedifferencewhenvary-ing   oneinputparameterfromitsminimumvalueto   itsmaximumvalue(Hamby,1994;Pannell,1997). 3.Resultsanddiscussion Duringthestudydays(33daystotalinthreeconsecutivevinegrowingseasons)theatmosphericconditionsof    thedrip-irrigatedMerlotvineyardwereingenerallydryandhotwithaverageTa,SR,RHandWs   of    19 ◦ C,   25.4MJm 2 day − 1 ,61.9%and1.1ms − 1 ,respectively(Table2).Dailyaverageof    VPDwasgenerallyaround1.0   kPaforthethreestudyseasons(Table2).Theaveragevaluesof     x  rangedbetween − 0.47and − 0.88MPa; − 0.42and − 1.0MPaand − 0.46and − 0.98MPa   duringthefirst,secondandthirdsea-sons,respectively.Thesevaluesindicatethat   Merlotvineyardwasnotunderseverewaterstressinanyof    thethreestudyperiods(WilliamsandTrout,2005;Sibilleetal.,2007).TheCHPVmethodtakesintoaccountthevariabilityof    thexyleminthecrosssectionthroughtheuseofthermocouplesto   differentdepth.Thus,in   orderto   exemplifythebehaviorofsapflowatthethreedifferentxylemdepth,6dayswereselectedintwo   pheno-logicalperiods:(i)   fruitsetand(ii)veraison(startofchangein   skincolorof    theberries)onthethreestudyseasons(Fig.2).Themax-imumsapfluxdensityvaluesfoundin   fruitsetperiodat5mmof depthin   thexylemwerearound80cmh − 1 ,   at10mmdepththemaximumvalueswerearound40cmh − 1 andat15mmdepththemaximumvalueswerearound18cmh − 1 (Fig.2a–c).Duringtheveraisonperiodsapfluxdensityvalueshadthesamebehaviorthatfruitsetperiodwithhighvaluesat5mmofdepthandlowvaluesat15mmdepth.Themaximumvaluesfoundat5mmof    depthinthexylemwerearound120cmh − 1 ,at10mmdepththemaximumvalueswerearound50cmh − 1 andat15mmdepththemaximumvalueswerearound20cmh − 1 (Fig.2d–f).Thesevaluesindicatethatxylemconductivitydecreasedwithdepthandtheexteriorside   of xylemwasveryyoungandactivecomparedwiththeinteriorside.Thiseffectis   consistentwithothersstudiescarriedoutin   olivestress(Fernándezetal.,2001;GiorioandGiorio,2003)   andvineyard(Intriglioloetal.,2009).Also,Fig.2showsthatthediurnaltrends of    threexylemdepths(5,10and15mm)   weresimilarthroughout  C.Poblete-Echeverríaetal./AgriculturalWaterManagement  109 (2012) 11–19 15  Table1 Statisticalparametersandcriteriaforevaluatingtheperformanceof    thecompensatedheat-pulsevelocity(CHPV)method.StatisticalparametersSymbolEquationOptimumMeanbiaserror MBE  1 N N   i = 1 ( E  i − O i )   0Meanabsoluteerror MAE  1 N N   i = 1  E  i  − O i   0Root   meansquareerrorRMSE   N i = 1 ( E  i − O i ) 2 N   0Index   ofagreement d 1 −   N i = 1 ( E  i − O i ) 2  N i = 1  E  i − O  +  O i − O  2   1where N    isthetotalnumberofobservations, E  i  and   O i  aretheestimated(Tsap)andobserved(Tr)values,respectivelyand O   isthemeanof    theobservedvalues.  Table2 Meandailyvaluesofairtemperature(Ta),relativehumidity(RH),solarradiation(SR),vaporpressuredeficit(VPD)andwindspeed(Ws)forthethreestudyseasons.Season SR(MJ   m − 2 day − 1 ) Ta( ◦ C)RH(%)   VPD(kPa)Ws   (ms − 1 )2006–200726.0( ± 6.5)17.5( ± 1.8)62.8( ± 9.6)0.93( ± 0.37)1.6( ± 0.5)2007–2008   26.6( ± 5.4)20.2( ± 2.5)61.0( ± 7.9)0.84( ± 0.13)0.7( ± 0.4)2008–200922.4( ± 1.1)19.0( ± 1.1)62.0( ± 7.7)1.04( ± 0.27)1.0( ± 0.5)Overall25.4( ± 6.0)19.0( ± 2.3)61.9( ± 8.2)0.86( ± 0.27)1.1( ± 0.6)Standarddeviationsvaluesinbrackets. theday.Sapflowrateswascharacterizedbyagradualincrementfrom8:00hreachinga   peakaround16:00handthengraduallydecreasingtominimumvaluesafter22:00h(LocaltimeGTM-4).Foranysapflowmethod,itis   necessaryto   convertanelec-tronicsignalinsapflow,whichinvolvestheuseof    empiricalvalues.Table3showstheresultsofthestatisticalanalysisof    the10woundsizespresentedbyGreenetal.(2003).   Thevaluesof    dailyvinetranspiration(mm   day − 1 )obtainedusingdifferentwoundsizeswerecomparedwiththereferencevaluesofvinetranspira-tion.Thesapflowtranspirationobtainedusingawoundsizeof 1.6mm   presentedanunderestimationof    16.7%,aRMSEvalueof 0.41mm   day − 1 (25%)anda   MAE   of0.32   mmday − 1 (19.4%).Ontheotherextreme,whensapflowtranspirationis   calculatedusingawoundsizeof    3.4mm,   itproducesanoverestimationof    38.6%withaRMSEvalueof0.73mmday − 1 (44.3%)anda   MAE   of    0.63mm   day − 1 (38.6%).Thebestagreementwasobtainedusingawoundsizeof2.4mm.   Inthiscasetheanalysispresenteda   MBEof    only − 0.04mm   day − 1 ( − 2.2%)a   RMSEvalueof    0.22mmday − 1 (11.9%),aMAE   of    0.18mm   day − 1 (8.8%)(Table3).Theseresultsareconsis-tentwiththestudypresentedbyBarretetal.(1995)who   indicatedthatthetotalwoundwidthis   likelyto   extendabout0.3mm   ateithersideof    thedrillhole.Thereforeawoundcorrectionof2.4mmseems 02040608010012014016000:00 04:00   08:00 12:00 16:00   20:00 00:00    S  a  p   f   l  u  x   d  e  n  s   i   t  y   (  c  m    h   -   1    ) Time (hh:mm) 5mm10mm15mmDOY348 Season 06-07Tsap = 1.81 mm day -1 02040608010012014016000:0004:0008:0012:0016:0020:0000:00    S  a  p   f   l  u  x   d  e  n  s   i   t  y   (  c  m    h   -   1    ) Time (hh:mm) 5mm10mm15mmDOY344 Season 07-08Tsap = 1.89 mm day -1 02040608010012014016000:0004:0008:0012:0016:0020:0000:00    S  a  p   f   l  u  x   d  e  n  s   i   t  y   (  c  m    h   -   1    ) Time (hh:mm) 5mm10mm15mm DOY351 Season 08-09Tsap = 1.93 mm day -1 02040608010012014016000:0004:0008:0012:0016:0020:0000:00    S  a  p   f   l  u  x   d  e  n  s   i   t  y   (  c  m    h   -   1    ) Time (hh:mm) 5mm10mm15mmDOY39 Season 06-07Tsap = 2.86 mm day -1 02040608010012014016000:00 04:00 08:00 12:00 16:00 20:00 00:00    S  a  p   f   l  u  x   d  e  n  s   i   t  y   (  c  m    h   -   1    ) Time (hh:mm) 5mm10mm15mmDOY36 Season 07-08Tsap = 2.61 mm day -1 02040608010012014016000:0004:0008:0012:0016:0020:0000:00    S  a  p   f   l  u  x   d  e  n  s   i   t  y   (  c  m    h   -   1    ) Time (hh:mm) 5mm10mm15mmDOY41 Season 08-09Tsap = 2.48 mm day -1 (a) (b) (e) (c) (f) (d) Fig.2. Exampleof    sapfluxdensityvaluesat5,   10and15mmdepthinfruitsetandveraisonperiodsforthethreevinegrowingseasons:Dayof    year(DOY)348(a),344(b)and   351duringfruitset,and39(d),36   (e)   and41   (f)duringveraison.
Search
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks